Mediterranean Tropical-Like Cyclones in Present and Future Climate

1OCTOBER 2014 C A V I C C H I A E T A L . 7493

LEONE CAVICCHIA Centro Euro-Mediterraneo sui Cambiamenti Climatici, Bologna, Italy

HANS VON STORCH Institute for Coastal Research, Helmholtz-Zentrum Geesthacht, Geesthacht, Germany

SILVIO GUALDI Centro Euro-Mediterraneo sui Cambiamenti Climatici, and Istituto Nazionale di Geoﬁsica e Vulcanologia, Bologna, Italy

(Manuscript received 9 May 2014, in ﬁnal form 15 July 2014)

ABSTRACT

The Mediterranean has been identiﬁed as one of the most responsive regions to climate change. It has been conjectured that one of the effects of a warmer climate could be to make the Mediterranean Sea prone to the formation of hurricanes. Already in the present climate regime, however, a few of the numerous low pressure systems that form in the area develop a dynamical evolution similar to the one of tropical cyclones. Even if their spatial extent is generally smaller and the life cycle shorter compared to tropical cyclones, such storms produce severe damage on the highly populated coastal areas surrounding the Mediterranean Sea. This study, based on the analysis of individual realistically simulated storms in homogeneous long-term and high-resolution data from multiple climate change scenarios, shows that the projected effect of climate change on Mediterranean tropical-like cyclones is decreased frequency and a tendency toward a moderate increase of intensity.

1. Introduction indicating that a few mesoscale vortices showing a striking resemblance to tropical latitude hurricanes It is known since the early times of meteorology occasionally do occur over the Mediterranean Sea. Such (Meteorological Office 1937) that the Mediterranean storms exhibit a cloud structure characterized by a central Sea, because of its peculiar features, such as the inter- eye surrounded by spiral-shaped cloud bands, and the play of the complex topography and orography and the strong winds associated can reach hurricane strength; geographical location between the subtropics and mid- the most intense storms of this kind have thus been latitudes, is characterized by the formation of a large termed Mediterranean hurricanes (medicanes). The lim- number of low pressure systems (Trigo and Davies 1999), ited available observations of the atmospheric fields mostly baroclinic in nature (Trigo et al. 2002). It has been around medicanes (Pytharoulis et al. 2000; Reale and suggested that the increased sea surface temperature in Atlas 2001; Moscatello et al. 2008; Lagouvardos et al. an anthropogenically warmed climate could lead to fa- 1999) showed further similarities between their dynamical vorable conditions for the formation of hurricanes in the features and the ones of tropical cyclones: a warm core, Mediterranean Sea (Gaertner et al. 2007). After the ad- a symmetric vertical profile, and rotating winds with vent of satellite meteorology in the last decades of the a well-defined structure around the cyclone eye. twentieth century, on the other hand, evidence emerged Theoretical (Emanuel 2005) and modeling (Davolio (Ernst and Matson 1983; Rasmussen and Zick 1987) et al. 2009; Fita et al. 2007; Miglietta et al. 2013) studies investigating the mechanisms driving the formation of medicanes pointed out that the genesis of such cyclones is Corresponding author address: Leone Cavicchia, Centro Euro- Mediterraneo sui Cambiamenti Climatici, Viale Aldo Moro 44, often characterized by the presence of cold air in the upper- 40127 Bologna, Italy. atmospheric layers, associated with a cutoff low over the E-mail: [email protected] Mediterranean area. Similar atmospheric configurations

DOI: 10.1175/JCLI-D-14-00339.1

Ó 2014 American Meteorological Society Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 7494 JOURNAL OF CLIMATE VOLUME 27 act as an important factor to enhance the atmospheric limited-area model the Consortium for Small-Scale instability to a level high enough to trigger the formation Modeling in Climate Mode (COSMO-CLM; Rockel of medicanes, which can occur above waters with surface et al. 2008). The model features nonhydrostatic equa- temperatures substantially lower than the ones of trop- tions and has 32 vertical sigma levels. Boundary and ical oceans (Miglietta et al. 2011; Palmén 1948). initial conditions are obtained by the NCEP–NCAR The small scale and low frequency of medicanes hinders reanalysis or by the global circulation model ECHAM5/ the systematic assessment of their properties in studies MPI-OM. The regional model is run at 0.098 (;10 km) exploiting data derived from reanalysis or long-term cli- horizontal resolution on a domain of 386 3 206 grid mate model experiments that have a too coarse spatial points covering the whole Mediterranean Sea and the resolution. Satellite-based detection, on the other hand, main mountain ranges around it. The model is run in a faces difficulties in reliably distinguishing such cyclones double-nested configuration, with an intermediate in- from baroclinic vortices. tegration on a 0.228 (;25 km) grid on a domain of 176 3 A method based on the dynamical downscaling of 106 grid points; only the final 10-km-resolution simula- global reanalysis data in a high-resolution atmospheric tion is used for the analysis. The atmospheric fields regional model, making use of the spectral nudging produced by the model are saved every hour, on 19 technique (von Storch et al. 2000), has been proven ef- pressure levels ranging between 1000 and 100 hPa at fective in successfully reproducing a number of histori- evenly spaced 50-hPa intervals. Spectral nudging (von cal medicane cases (Cavicchia and von Storch 2012). Storch et al. 2000) is applied on the horizontal compo- Such a method has been used in Cavicchia et al. (2014) nents of the wind field components for spatial scales to reconstruct the climatology of medicanes over the larger than approximately 4 times the grid size of the period 1948–2010. forcing fields and on vertical levels above 850 hPa, while In the present work we extend the aforementioned the small-scale components and the lower vertical levels dynamical downscaling method to study the properties are let free to evolve following the small-scale dynamics of all Mediterranean tropical-like cyclones (MTLCs) and high-resolution orography. including, besides the most intense medicanes, weaker All the storms exhibiting tropical-like features tropical-like storms. The same method is then applied are detected in the four downscaled simulations, using to future climate scenarios, to estimate the response to an objective detection algorithm specifically designed climate change of the frequency and intensity of Medi- for MTLCs. The geographical distribution, seasonal terranean tropical-like cyclones. cycle, and intensity distribution of the detected storms are then compared. The detection algorithm consists of 2. Methods the following steps. The sea level pressure minima with a gradient larger than 20 Pa over three grid points are The climatological properties of MTLCs over the last assigned to the same track if they lie within a radius of six decades are assessed by downscaling the National 100 km at two consecutive hourly time steps. The tracks Centers for Environmental Prediction–National Center composed of at least six points are kept for further for Atmospheric Research (NCEP–NCAR) reanalysis analysis if the following occurs: no more than half of data (Kalnay et al. 1996). To estimate the effect of cli- the positions correspond to land-covered model grid mate change, we performed the downscaling of two dif- points, the two most distant positions are separated by ferent 30-yr (2070–99) time slices of climate scenarios at least 200 km, and a change in the propagation di- produced by the global circulation model ECHAM5/Max rection does not exceed p/3 for more than half of the Planck Institute Ocean Model (MPI-OM) (Roeckner positions. A cyclone is classified as an MTLC if for et al. 2003) forced by the greenhouse gas concentrations more than 10% of the track or more than 6 h the fol- prescribed in the Intergovernmental Panel on Climate lowing occurs: it is positioned in the symmetric and Change (IPCC) Special Report on Emissions Scenarios warm core sector of the phase space defined in Hart (SRES) for emissions scenarios A2 and B1. To minimize (2003) with a radius of 100 km, the 10-m wind speed the effect of systematic biases of the global model chosen, within a 50-km radius around the pressure minimum 2 we also performed the downscaling of a 30-yr (1960–89) exceeds 18 m s 1, and its average is larger at 850 hPa time slice of a present climate simulation produced by the than at 300 hPa. A sensitivity analysis to the numerical same ECHAM5/MPI-OM global circulation model, values of the different parameters entering the detection forced by twentieth-century greenhouse gases concen- algorithm, and the validation and optimization pro- trations (C20 simulation); the scenario simulations are cedure employed to determine the thresholds described thus compared with the control run. The dynamical above, have been presented and discussed in a previous downscaling of atmospheric fields is performed using the work (Cavicchia et al. 2014).

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 1OCTOBER 2014 C A V I C C H I A E T A L . 7495

FIG. 1. Genesis density of MLTCs (ﬁrst location in the track) per 18318 box per decade: (a) NCEP, (b) C20, (c) B1, and (d) A2.

3. Results a consistent formation pattern, with most of the events produced in the western Mediterranean and a secondary a. Statistical properties of MTLCs maximum in the Ionian Sea; a few minor differences in The locations of formation (Fig. 1) of all the MTLCs the exact location of the cyclogenesis maxima are, how- detected in the reanalysis-driven downscaling show ever, visible. On the other hand, the overall frequency of a pattern characterized by two maxima located in the detected MTLCs in the C20 simulation is about 20% western Mediterranean and in the Ionian Sea, and very lower compared to the one found in the NCEP down- low activity in the eastern Mediterranean. Such a for- scaling. Such discrepancy is due to the bias in the atmo- mation pattern is similar to the one found in Cavicchia spheric variables on which MTLC formation depends, et al. (2014) for the subset of the stronger medicanes;1 mostly the vertical temperature gradient controlling the on the other hand, including the full set of MTLCs the atmospheric stability. The role of vertical temperature frequency is increased about 3 times with respect to the gradients in the formation of Mediterranean tropical-like one found for medicanes. A comparison with the limit- cyclones is discussed in detail in section 3b below; it ed satellite-based observations available (Fita 2014) is, however, anticipated that the C20 simulation shows a and modeling studies (Walsh et al. 2014; Romero and negative bias in the amplitude of the temperature dif- Emanuel 2013) using different methodology and data ference between the high troposphere and sea surface shows similar patterns. The C20 simulation produces ranging between 18 and 28C with respect to the NCEP- driven simulation in the areas of MTLC formation (see Fig. 5e). In the B1 simulation, the MTLC formation rate shows a 1 Following Cavicchia et al. (2014), MTLCs are classiﬁed as moderate reduction (about 15%) compared to C20. A medicanes if the track fulﬁlls the additional criterion to have 21 few small-scale changes in the locations of formation are a value of the peak wind speed VMAX larger than 29 m s for at least 4 h. visible (e.g., the maximum west of Sardinia); the decrease

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 7496 JOURNAL OF CLIMATE VOLUME 27

TABLE 1. Total number of cyclones detected in the main subregions in the different simulations.

NCEP C20 B1 A2 Western (lon , 98E) 124 48 43 23 Ionian (lat , 40.58N; 78 27 23 6 168E , lon , 22.58E) Eastern (lon . 22.58E) 27 8 6 4 of the formation rate, however, is equal over the main different subregions (Table 1). In the A2 simulation, the decrease in frequency becomes substantial (about a 60% decrease with respect to C20). Moreover, the frequency of formation in the Ionian Sea area shows a strong decrease. The seasonal cycle of cyclone formation in the four different experiments characterizes Mediterranean tropical- like storms as a cold season phenomenon (Fig. 2). The FIG. 2. Average number of MTLCs per month per decade (bars) seasonal cycle, like the formation pattern, shows fea- and 3-month running mean (lines) in NCEP (blue), C20 (cyan), B1 tures similar to the ones found for medicanes only. The (yellow), and A2 (red). relative difference between the four simulations reflects, except for a number of minor fluctuations, the changes where RH represents relative humidity and the sum is in the respective overall frequency. A noticeable shrink- calculated on the model output levels, evenly spaced at ing of the formation season, however, emerges in the A2 50-hPa intervals, and the wind shear between the 850- and 2 2 1/2 simulation. 250-hPa levels WS 5 [(u250 2 u850) 1 (y250 2 y850) ] . The intensity distribution of MTLCs (Fig. 3) is derived Adapting the approach used in Cavicchia et al. (2014) for using as an intensity metric the maximum value in a 50-km the subset of most intense medicanes to the full set of radius around the pressure minimum of the VMAX, dur- Mediterranean tropical-like cyclones, we find that the ing each storm lifetime. This variable has been shown, detected MTLCs form when DT . 578C, HI . 1000, and 2 through validation with observations, to provide a good WS , 18 m s 1. We analyze, for each of the environmental criterion to discriminate medicanes from weaker Medi- variables WS, HI, and DT, the changes in future climate terranean tropical-like storms (Cavicchia et al. 2014). simulations of the mean state and of the percentage of Both the mean value of the intensity distribution and days in which the threshold corresponding to MTLC for- the percentile corresponding to storms exceeding the mation is exceeded. We consider only the cold season, 21 medicane intensity threshold (29 m s ) show (Table 2) averaging the model variables over the months from an increase in the climate change simulations compared October to March (ONDJFM). The exceedance prob- to the control run. Moreover, events reaching extreme abilities are obtained from the model daily mean fields, 21 (*50 m s ) wind speeds are present in both scenario calculating for every grid point the quantity P(V) 5 simulations, while the most intense event in the control 100 ND(V . T )/NDTOT, where ND is the number of 21 simulation has a maximum peak wind speed of 43 m s . days in which the variable V has a value larger than the T b. Environmental factors associated with decreased threshold value and NDTOT is the total number of frequency of storms days in the simulation. For the changes in the mean state, we consider the relative difference 100(VB1,A2 2 To investigate the decreased frequency of cyclones VC20)/VC20. detected in the scenario simulations, we analyzed a num- The frequency of all environmental configurations fa- ber of factors that have been shown in the literature vorable for MTLC formation shows a coherent decrease (Cavicchia et al. 2014; Tous et al. 2013) to be related to in both future climate scenarios. The magnitude and pat- the formation of MTLCs. Such factors include the verti- tern of the change of the three analyzed factors is different. cal temperature gradient between the sea surface and the The cold season mean state of WS in the C20 simula- 2 high troposphere DT 5 jT(350 hPa) 2 SSTj, the vertically tion ranges between 22 and 24 m s 1 in the western part of integrated humidity content of the atmosphere the basin; in the eastern Mediterranean, it shows larger 2 values, up to 36 m s 1, with a southeastward gradient. The 1000 hPa HI 5 å RH(l), mean state of the wind shear during the cold season is l5100 hPa increased in warmer climates. The magnitude of the

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 1OCTOBER 2014 C A V I C C H I A E T A L . 7497

FIG. 3. The maximum value of VMAX detected in a 50-km radius during the storm lifetime for (a) NCEP, (b) C20, (c) B1, (d) A2. Bins 21 21 represent 3 m s intervals centered at the values indicated on the x axis. The lower bound at around 25 m s in VMAX is due to the 2 18 m s 1 cutoff applied to the average wind speed in the detection algorithm (see section 2). relative increase is, in the areas of MTLC formation, 3%– The cold season mean state of the integrated relative 5% in the B1 scenario (not shown) and 5%–10% in the humidity is characterized in the C20 simulation by a A2 scenario (Fig. 4a), while it remains unchanged in the strong meridional gradient, being about 3 times lower in southern part of the basin. The frequency of conﬁgura- the southern Mediterranean compared to the northern tions with wind shear low enough for MTLC formation part of the basin. The HI mean state shows a moderate 2 (WS , 18 m s 1) shows a moderate decrease in the same relative decrease in the areas of MTLC formation of area of about 2% in the B1 scenario (not shown) and 4% 3% in B1 (not shown) and 5% in A2 (Fig. 4c), with the in the A2 scenario (Fig. 4b). larger difference in the southeastern Mediterranean.

TABLE 2. Summary of the main statistical properties of detected MTLCs.

NCEP C20 B1 A2 Sample size 317 127 105 52 Frequency (events per year) 3.91 6 1.85 3.26 6 2.02 2.68 6 1.33 1.35 6 1.17 2 Mean intensity (m s 1) 31.01 6 4.09 30.79 6 3.83 31.77 6 5.40 31.13 6 4.25 2 29 m s 1 intensity percentile 63.09 63.78 63.81 67.32

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 7498 JOURNAL OF CLIMATE VOLUME 27

FIG. 4. Changes in large-scale parameters related to MTLC formation in future climate scenarios. (a) Mean value of wind shear in the 2 cold season (ONDJFM), relative difference between the A2 and C20 scenarios. (b) Frequency of days with WS , 18 m s 1, difference between the A2 and C20 scenarios. (c) As in (a), but for relative humidity. (d) Frequency of days with HI . 1000, difference between the A2 and C20 scenarios.

The frequency of configurations with atmospheric hu- associated with the formation of medicanes. It is worth midity high enough for MTLC formation (HI . 1000) disentangling the effect of the two remaining factors—the shows a comparable decrease in the scenario simulations, mean value of the temperature difference DT and the of about 3% in B1 (not shown) and 5% in A2 (Fig. 4d). frequency of cold anomalies T~ in the high troposphere— The major change between present and future climate whose combination contributes to reaching the value of is found in the vertical atmospheric temperature gradi- instability needed for the formation of MTLCs, since they ent. The probability of atmospheric configurations with can be influenced differently by climate change. The mean DT high enough for MTLC formation (DT . 578C) value DT decreases about 28C, or 3%, in B1 (Fig. 5a)and shows a decrease in the central Mediterranean of about 38C, or 5%, in A2 (Fig. 5c), with a larger difference in the 13%–15% in the B1 scenario (Fig. 5b) and of 19%–21% central part of the basin. Such a decrease reflects the fact in the A2 scenario (Fig. 5d). The realization of a given that the atmosphere responds faster than ocean to climate value of the vertical temperature gradient DT can be warming (Collins 2014). The increase of the mean value of ~ written as DT 5 DT 2 T~ 1 S, where DT represents the sea surface temperature in the A2 scenario is shown in mean state of the temperature gradient, T~ 5 T 2 T is the Fig. 5f. The frequency of cold anomalies T~ of the 350-hPa deviation of the 350-hPa temperature from its mean state, temperature with respect to its mean value (Fig. 5g), on and S~5 S 2 S is the deviation of the sea surface tem- the other hand, shows a slight increase in the scenario perature from its mean state. It was found in Cavicchia simulations, for a fixed amount of the anomaly value. et al. (2014) that the formation of medicanes is driven by Since the mean value has changed, however, reaching the factor T~,whereasS~ does not show a coherent signal the threshold associated with MTLC formation requires

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 1OCTOBER 2014 C A V I C C H I A E T A L . 7499

FIG. 5. Changes in the atmospheric temperature and MTLC formation in future climate scenarios. (a) Mean value of DT in the cold season (ONDJFM), relative difference between the B1 and C20 scenarios. (b) Frequency of days with DT . 578C, difference between the B1 and C20 scenarios. (c),(d) As in (a),(b), but for the A2 and C20 scenario difference.(e) Mean value of DT in the cold season, difference between the C20 and NCEP simulations. (f) Mean value of SST in the cold season, difference between the A2 and C20 scenarios. (g) Frequency of days (basin average, sea points only) with anomalies (with respect to the monthly climatological value) of T (350 hPa) larger than the value indicated on the x axis in C20 (blue), B1 (green), and A2 (red). (h) Mean value of DT (basin average, sea points only) in the cold season in different CMIP3 models and downscaled simulations in C20 (stars), B1 (diamonds), and A2 (circles).

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 7500 JOURNAL OF CLIMATE VOLUME 27 a larger anomaly that is less frequent, and this explains Acknowledgments. We acknowledge the modeling the decreased probability of occurrence of values of DT groups the Program for Climate Model Diagnosis and above the threshold. The increased frequency of anom- Intercomparison (PCMDI) and the WCRP’s Working alies of equal amplitude in the scenario simulations is Group on Coupled Modelling (WGCM) for their roles in coherent with the expectation of a slower warming of the making available the WCRP CMIP3 multimodel data- upper troposphere at high latitudes (Collins 2014), since set. Support of this dataset is provided by the Office of cold anomalies are associated with the intrusion in the Science, U.S. Department of Energy. L. C. was partially Mediterranean region of air masses originated in higher supported by the NextData project, funded by the Italian latitudes; the proof of such conjecture, however, goes Ministry of Education, Universities and Research. beyond the scope of the present work. Summarizing the main results of the discussion, we find that (i) in a warmed climate, the mean temperature difference between the REFERENCES sea surface and the high troposphere is decreased, due to Cavicchia, L., and H. von Storch, 2012: The simulation of medicanes the atmosphere warming faster than the ocean; and in a high-resolution regional climate model. Climate Dyn., 39, (ii) the frequency of negative fluctuations of a given mag- 2273–2290, doi:10.1007/s00382-011-1220-0. ——, ——, and S. Gualdi, 2014: A long-term climatology nitude of the high troposphere temperature around the of medicanes. Climate Dyn., 43, 1183–1195, doi:10.1007/ mean value remains mostly unchanged in a warmed cli- s00382-013-1893-7. mate; however, because of the decrease of the mean Collins, M., and Coauthors, 2014: Long-term climate change: Pro- value, a larger anomaly is required to reach the threshold jections, commitments and irreversibility. Climate Change 2013: associated with cyclone formation, and this explains the The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 1029–1136. [Available online at http://www. reduced frequency of storms. climatechange2013.org/images/report/WG1AR5_Chapter12_ Finally, we compared the changes in the mean value of FINAL.pdf.] DT obtained in our simulations with the values obtained Davolio, S., M. M. Miglietta, A. Moscatello, F. Pacifico, A. Buzzi, and in a number of models from phase 3 of the Coupled R. Rotunno, 2009: Numerical forecast and analysis of a tropical- 9, Model Intercomparison Project (CMIP3; Meehl et al. like cyclone in the Ionian Sea. Nat. Hazards Earth Syst. Sci., D 551–562, doi:10.5194/nhess-9-551-2009. 2007). The spread in the value of T (Fig. 5h) reflects the Emanuel, K., 2005: Genesis and maintenance of ‘‘Mediter- different model biases, but the strength of the decrease ranean hurricanes.’’ Adv. Geosci., 2, 217–220, doi:10.5194/ found in the B1 and A2 scenarios is comparable in most adgeo-2-217-2005. of the models, increasing the confidence in the robust- Ernst, J. A., and M. Matson, 1983: A Mediterranean tropical storm? 38, ness of the projected changes in storm frequency. Weather, 332–337, doi:10.1002/j.1477-8696.1983.tb04818.x. Fita, L., cited 2014: Medicanes: Mediterranean tropical-like storms. [Available online at http://www.uib.es/depart/dfs/meteorologia/ 4. Conclusions and outlook METEOROLOGIA/MEDICANES.] ——, R. Romero, A. Luque, K. Emanuel, and C. Ramis, 2007: Storms exhibiting dynamical features similar to the ones Analysis of the environments of seven Mediterranean tropical- of tropical cyclones occur occasionally over the Mediter- like storms using an axisymmetric, nonhydrostatic, cloud re- solving model. Nat. Hazards Earth Syst. Sci., 7, 41–56, ranean Sea. For a fraction of such storms, the tropical-like doi:10.5194/nhess-7-41-2007. intensification is strong and long-lasting enough to allow Gaertner, M., D. Jacob, V. Gil, M. Domínguez, E. Padorno, them to attain hurricane strength. Different from tropical- E. Sánchez, and M. Castro, 2007: Tropical cyclones over the latitude hurricanes, cold air aloft is required in order to Mediterranean Sea in climate change simulations. Geophys. 34, enhance thermodynamic disequilibrium and trigger the Res. Lett., L14711, doi:10.1029/2007GL029977. Hart, R. E., 2003: A cyclone phase space derived from thermal formation of Mediterranean tropical-like cyclones. wind and thermal asymmetry. Mon. Wea. Rev., 131, 585–616, The formation of MTLCs is found to occur mostly in doi:10.1175/1520-0493(2003)131,0585:ACPSDF.2.0.CO;2. the central and western parts of the basin during the Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Re- cold season. The frequency of MTLCs is projected to analysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, , . decrease in future climate scenarios, as a consequence of doi:10.1175/1520-0477(1996)077 0437:TNYRP 2.0.CO;2. Lagouvardos, K., V. Kotroni, S. Nickovic, D. Jovic, G. Kallos, the lower frequency of environmental configurations fa- and C. J. Tremback, 1999: Observations and model simu- vorable for their formation. The projected intensity of lations of a winter sub-synoptic vortex over the central simulated MTLCs in scenario simulations shows on the Mediterranean. Meteor. Appl., 6, 371–383, doi:10.1017/ other hand a tendency to a moderate increase. The robust S1350482799001309. assessment of changes in MTLC intensity requires, Meehl, G. A., C. Covey, K. E. Taylor, T. Delworth, R. J. Stouffer, M. Latif, B. McAvaney, and J. F. B. Mitchell, 2007: The however, further investigation based on a larger sample WCRP CMIP3 multimodel dataset: A new era in climate of simulated events, and on the analysis of the physical change research. Bull. Amer. Meteor. Soc., 88, 1383–1394, mechanisms involved in MTLC intensification. doi:10.1175/BAMS-88-9-1383.

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC 1OCTOBER 2014 C A V I C C H I A E T A L . 7501

Meteorological Office, 1937: Weather in the Mediterranean. Vol. 1, Rockel, B., A. Will, and A. Hense, 2008: The regional climate General Meteorology, Her Majesty’s Stationery Office, 218 pp. model COSMO-CLM (CCLM). Meteor. Z., 17, 347–348, Miglietta, M. M., A. Moscatello, D. Conte, G. Mannarini, doi:10.1127/0941-2948/2008/0309. G. Lacorata, and R. Rotunno, 2011: Numerical analysis of a Roeckner, E., and Coauthors, 2003: The atmospheric general cir- Mediterranean ‘hurricane’ over southeastern Italy: Sensitivity culation model ECHAM 5. Part I: Model description. MPI experiments to sea surface temperature. Atmos. Res., 101, Rep. 349, 140 pp. [Available online at http://www.mpimet. 412–426, doi:10.1016/j.atmosres.2011.04.006. mpg.de/fileadmin/publikationen/Reports/max_scirep_349.pdf.] ——, S. Laviola, A. Malvaldi, D. Conte, V. Levizzani, and C. Price, Romero, R., and K. Emanuel, 2013: Medicane risk in a changing 2013: Analysis of tropical-like cyclones over the Mediterra- climate. J. Geophys. Res. Atmos., 118, 5992–6001, doi:10.1002/ nean Sea through a combined modelling and satellite approach. jgrd.50475. Geophys. Res. Lett., 40, 2400–2405, doi:10.1002/grl.50432. Tous, M., R. Romero, and C. Ramis, 2013: Surface heat fluxes in- Moscatello, A., M. M. Miglietta, and R. Rotunno, 2008: Observa- fluence on medicane trajectories and intensification. Atmos. tional analysis of a Mediterranean ‘hurricane’ over south- Res., 123, 400–411, doi:10.1016/j.atmosres.2012.05.022. eastern Italy. Weather, 63, 306–311, doi:10.1002/wea.231. Trigo, I. F., and T. D. Davies, 1999: Objective climatology of cy- Palmén, E., 1948: On the formation and structure of tropical hur- clones in the Mediterranean region. J. Climate, 12, 1685–1696, ricanes. Geophysica, 3, 26–38. doi:10.1175/1520-0442(1999)012,1685:OCOCIT.2.0.CO;2. Pytharoulis, I., G. Craig, and S. Ballard, 2000: The hurricane-like ——, ——, and G. R. Bigg, 2002: Climatology of cyclogenesis Mediterranean cyclone of January 1995. Meteor. Appl., 7, 261– mechanisms in the Mediterranean. Mon. Wea. Rev., 130, 549– 279, doi:10.1017/S1350482700001511. 569, doi:10.1175/1520-0493(2002)130,0549:COCMIT.2.0.CO;2. Rasmussen, E., and C. Zick, 1987: A subsynoptic vortex over the von Storch, H., H. Langenberg, and F. Feser, 2000: A spectral nudging Mediterranean with some resemblance to polar lows. Tellus, technique for dynamical downscaling purposes. Mon. Wea. 39A, 408–425, doi:10.1111/j.1600-0870.1987.tb00318.x. Rev., 128, 3664–3673, doi:10.1175/1520-0493(2000)128,3664: Reale, O., and R. Atlas, 2001: Tropical cyclone–like vortices in the ASNTFD.2.0.CO;2. extratropics: Observational evidence and synoptic analysis. Wea. Walsh, K., F. Giorgi, and E. Coppola, 2014: Mediterranean warm- Forecasting, 16, 7–34, doi:10.1175/1520-0434(2001)016,0007: core cyclones in a warmer world. Climate Dyn., 42, 1053–1066, TCLVIT.2.0.CO;2. doi:10.1007/s00382-013-1723-y.

Unauthenticated | Downloaded 09/29/21 08:14 PM UTC